Power of AI in Cybersecurity: A Conversation with Jason Dion

Hello everyone, and thank you for tuning back into another episode of the AI Automations Podcast. And today we have a really special guest with us here today, Jason Dionne, who is a dear friend of mine. We have been friends for 6, 7, 8 years. I forgot, I lost track. But anyway, Jason Dionne is the Chief Product Officer at Accolade, and he is also the lead instructor at Dionne Training Solutions. He has helped over 1 million people around the world take and pass their IT, cybersecurity, and project management certifications. Definitely a pro in all of those fields. And these days he's focused on creating his own certifications in the cybersecurity space. And, and he is also here to help those of you who want to learn the intersection between cybersecurity and AI. And yeah, thank you, Jason, for being here on the podcast.

Thanks for having me. I'm happy to be here.

Perfect. All right. So as we start off here, can you just talk to us a little bit about how you see AI transforming the field of cybersecurity in the near future?

Yeah, definitely. You know, AI has been the real buzzword these last, you know, 12 to 18 months, I would say. And before that, it was blockchain and crypto. And it seems there's always something in the field that throws everybody for a loop when they come out and be like, oh my God, it's the end of cybersecurity as we know it because AI is going to take over and we're all going to lose our jobs. And that's the big thing I've been hearing a lot. And really, my my belief is that is not the case, right? When it comes down to it, AI is an assistance tool. It can help offload a lot of tasks that are repetitive, just like automation could. We went from automation to machine learning to now AI. And you know, if you go back to the '80s and '90s, people all thought that everyone was going to lose their jobs because automation was happening, right? And obviously that didn't happen. People adapt, people move on, and we start taking away lower level and repeatable processes and start moving those into systems like AI, machine learning, or automation. And then we as humans evolve to take on. Those higher-level roles. Now, AI is a little bit different, as you know. It does have a much higher capability to do things, but it is still not perfect. And, you know, you fast forward 5 or 10 years, it's going to be doing a lot more things that we currently do as blue-collar professionals or white-collar professionals, excuse me. So if you're in finance, if you're in accounting, if you're a lawyer, if you're a radiologist, AI has applications in all of these areas. And so it will have an impact on the workforce. But I think what's important is for us as employees and us as people in our careers is to identify Where are the areas that we can shift into that are not replaceable? Where are the places that they need our brains and our intelligence beyond what AI can do? In the world of cybersecurity, I think the biggest impact is going to be helping us in the triage area. So when something is happening on the network, AI can detect that, AI can see it, they can classify it, and they may even be able to respond to it. But I will tell you, as somebody who worked on a national and worldwide level in cybersecurity, one of the last networks I worked on, we had over 1 million endpoints located around the world. We had the ability to turn on AI to do the classification and response at that time. We had chosen not to do that. And the reason is sometimes AI will make a decision that doesn't take into account the whole picture. And when we talk about businesses, for instance, let's say your website started having some people doing a ping scan across your server to see, is it open and running and is certain ports on your server open? Well, we can see that and the AI may go, oh, this is bad. It looks like an attack. Let me shut down my server and that way nobody will get to it. Yes, that protects your server, but guess what? You just closed down your entire business because now you can't take credit cards. Now you can't fulfill orders. And so it has to understand where are the actions it can take. And a lot of times that is a risk management decision being made by humans to determine, do I want to allow the attack to continue for another hour or 2 so that I can keep doing the sale, or do I take down my site so I can fix it and lose all those potential sales? And this is something that AI is still not great at doing. So that's kind of my initial thoughts on AI as we, as we talk. talking about is there's great use cases for it, and there's also places where we still want a human in the loop because we just don't trust it yet.

Hmm, that's a very interesting approach because everyone's been talking about like AI adoption, AI adoption, you gotta get on. Um, so I guess that kind of leads me to ask you, how do you approach and what, what kind of frameworks or, um, things would you think about when it comes to balancing AI and automations with making human decisions in, you know, particularly in cybersecurity?

Yeah, so one of the things that I personally don't like about AI versus automation is that with automation, we gave it very cool, very clear rules, right? We know exactly what the inputs are, and it follows very well-defined paths. For example, if you're a marketer, you could say when somebody goes to my website and enters their email, it can start doing these functions. And when they get here, there's a decision point. And if they bought this, it goes this way. And if they didn't, it goes this way. And you can really follow that path very clearly, know exactly what's happening. AI, not so much, right? It's a black box. What goes in is some input from your customer. What comes out is the response, and you don't know what's happening in between. Now, that's great when the systems are programmed well and they've learned over time and have a big dataset. But the problem is, if it doesn't have a really good dataset, you don't know when it's making a mistake. You don't know when it's making racist decisions, right? Or decisions that are going to hurt your business. For instance, I read a report. There was a Police system here in Florida that started using AI as a way to give a rating on a— during adjudication. Like, let's say, you know, somebody picked up Helena because she walked out of Walmart and she had a stick of gum in her pocket she didn't pay for, right? She would go in front of a judge and say, hey, you were caught stealing. You stole that $0.25 pack of gum. How do you plead? And you'll say guilty or not guilty. And based on that, the judge will then determine, do we just give you a slap on the wrist? Do we give you a fine? Do we put you in jail for the night? Do we put you in jail for a long period of time? Do we take away your passport? How serious is this? Well, they had an AI system help make these determinations to give the judge a recommendation. And what they found was that every time it was a Black person in front of the judge, they got put into custody and were not put out on bail. When it was a white person, more often than not, they were allowed out on bail. Now, why is that? Well, because the system was trained on the last 20 years of experience that it had from all the court cases that judges had made. And people tend to be racist, right? They have these racial overtones because Black people were incarcerated a lot more than white people. And so when the AI started making decisions, it started doing that exact same thing. I read another study when they were trying to train some machine machine learning and AI algorithms on determining what was a party and what wasn't. And they showed the system thousands of pictures. And when they started, they started showing pictures and said, okay, these are pictures of a party and these are not. And so if you saw a bunch of smiling people around a table with food, they considered that a party. Uh, if they saw, you know, a bunch of people sitting around a table at work doing work, it was not a party. Well, one of the things they started finding was that anytime there was an African American or a Black person in the picture, it was defined as not a party. And we're like, why is that? And when they look back at the data they gave it, the 5,000 images they gave it, none of them had Black people in it. And so the system became racist and said, well, if somebody is Black, it must not be a party. And that was just because of a bad data input. And so when you're training your algorithms in AI, it really does matter how you train it because you can create these biases that we have in the real world and they carry over into our systems. And this affects us when we're using this for AI, for training, for automations, and all those other kinds of things, because all of this carries over based on the dataset you got.

Wow. Wow, that's really, really interesting what you just said. Yeah, I've definitely seen similar cases. I remember I watched a video a couple of days ago when they were saying that when they were trying to build recruiting systems using AI, they ran into the same kind of biases because they trained the AI on, you know, what a good candidate looks like and what a bad candidate looks like. And except One of the things that they overlooked was when they were training on the good candidates, a lot less of them were Black. So the system unfairly thought that people of color were just not as qualified candidates, which is not the case.

And it's completely not an AI issue, right? It's a training model issue. And if you go in the regular world without computers, the same type of biases happen. If you look over the last 20 years and you look at Silicon Valley funding for startups, right? You can have a startup, Helena, and I'll have a startup, right? Right?

Yeah.

Um, you happen to be an Asian woman in her 30s. I am a white guy in my 40s. If we both walk into an investor and we pitch the exact same thing, I am 20 to 50 times more likely to get funded than you are. Uh, women in tech, generally it's 1 to 3% of venture startups are getting funded if they are women-led, uh, whereas it's a white male, it's like 95%. Um, Black founders also very heavily discriminated against. And it's not because they're like, oh, you're, you're, you're Asian or you're Black, and that's the reason we're not funding you. It's they start looking like, well, over the last 20 years, who are our successful startups? Well, Zuckerberg, he's a white guy in his 30s, right? Uh, you know, Travis, uh, from Uber, right? Billion-dollar company. Uh, wasn't a great founder, but, but, you know, again, white guy, right? But if you look at all these founders, they— most of them are white folks, right? And, uh, and white males specifically. And so the, the guys who are writing the checks go, oh, you know what, I'm used to seeing white guys successful. Uh, Black folks aren't, or Asian folks aren't. And it's not that they aren't, it's that they're not getting a chance, right? You know, 1 out of 100 are getting a chance versus 99 out of 100 for the white folks. Right. And so the system is just train this discrimination and this racism inside the system, and that carries over to our AI models too. So we have to be aware of those things.

Yeah, yeah. Not to make this an ethical debate about how to train AI models, so, and not to make this super focused on race, but how do we— how would you think that we can overcome some of the challenges when it comes to training AI models so that they become more fair and they can make better decisions?

Well, I think one of the big things is you have to be aware that the bias exists, right? And you have to be very careful of the training data you put into it. I have a friend who owns a company that works with startups to sell their companies to private acquisition firms when they get large enough and they want to do it. They get a lot of people say, hey, you know, Helena's got this AI company and she wants somebody to buy her. He probably gets, you know, thousands of those a month, people saying, hey, please represent me. His team, they have enough capacity to do about 5 to 10 deals per year. So they're highly selective. And so what they did was actually built an AI model to determine determine which deals should we actually look at with a human versus which ones should we just throw away and say, eh, not, not good enough, right? And so they've created this whole AI model, and they were very particular in going back through the last 20 years of history of the types of companies they want to work with, the types of founders they want to work with, the amount of money they want to work with. Because if you're a $500,000 a year business, you're too small to make a dent for them. If you're a $500 million a year business, you're too big and they can't represent you. So there's like a sweet spot for that. And so as they've done that, they've collected data. They have about 2,000 data points, and they have They have the system go through and the system classified it. And then as the system went through, they went back and said, I agree or I don't agree on each one of those 2,000. Now that they have a good secure baseline that does not have the biases inside of it because they've, they've used a large dataset that they've cleaned up, it is able to take care of that issue. So, so it is something you can do. But the thing is, I've seen a lot of companies when it comes to AI, they're saying we have AI, right? That's been the big buzzword in the last 3 to 6 months. Everybody has AI, right? But what are they all doing? Pretty much, I would say 90% of them out there are using an API call to OpenAI to use ChatGPT. on the backend, and they are, quote, doing AI now because they're sending your prompt to there, taking it back, and then popping it in your software. The problem is it's a black box. You don't know what it was trained on because a GPT is a generic model. It's trained on everything. It's trained on the entirety of the internet. And we all know the internet is just a wonderful source of truth on everywhere you go. Not. And it's not a complete racist, you know, um, hellhole. It is, right? Um, and so these are the kind of problems that we have to deal with when you're dealing with a generic model. If you have your own AI model that you've built, And you've trained it, you can then be more confident in what it's going to do. But when you just say, go out and listen to Twitter for 24 hours, and that's your training, which Microsoft did back in 2016, 2017 with Tay, which was their Twitter service that was using AI, within 24 hours, it started spewing out the most racist, misogynistic stuff you've ever heard in your life, because all it was doing was reading what was on Twitter and then responding that way. And so these things are self-training. They are self-learning. And the more data they get, the more they train. But if they get bad data or racist data or misogynistic data, misogynistic data or whatever, it's going to give you bad information. And I know when I use ChatGPT, for instance, um, you know, I'll prompt it and have it help me write practice exam questions or scripts or whatever. You have to double-check what it gives you because it will tell you it is right and it thinks it's right, but it's not. Like, half the time it's wrong. And ChatGPT-4 is much better than 3.5, and 3.5 is better than 3, and it'll get better over time. But again, it's just a function of how much data it's ingesting. And the more users that are seeing it, the more they're saying this was a good response and this was a bad response. the more accurate it'll be over time. But if you have a very, you know, model trained on 5 or 10 data points, it's going to be vastly wrong, right? Because if you have a 10,000 or 100,000-point dataset, it's going to be much more clear, especially if you have a diverse dataset to begin with.

Right, right, right, right. Perfect. So, so a lot of people associate like AI with text thanks to ChatGPT, and they think that AI is only going to be used for text.

Oh, heck no.

Yeah. So what you said earlier kind of got my brain turning a bit. You were saying that it's kind of like, oh, in cybersecurity more specifically, it's used to classify whether something is a threat or not. So can you kind of talk about the other use cases for AI beyond just ChatGPT and particularly how it relates in the field of cybersecurity?

Yeah, definitely. So I, I think the first thing to say is that, you know, your audience is familiar with AI, so they're probably already familiar with this. But, you know, there is— AI is the big umbrella term, but underneath AI you also have machine learning and big data. Those all go together, right? Um, and generally machine learning is kind of dumb AI, and under that you even have dumber, which is automations, which is just follow this 7-step process where machine learning is really good for categorization. So we use that very heavily inside of cybersecurity. And then when you start doing AI, you have a system that can think for itself and make decisions based on the data it's analyzing. So it's just a much more advanced automation that can actually think and create. Other areas that we see it used in, and myself, we use a lot of these different areas in my training business. We actually use AI voices to correct issues that our instructors made during the course. So for example, when I'm teaching something like Security+, it's a 30 or 40-hour course, which means I am talking for 80 to 100 hours, and then my team edits it down to the 30 or 40 hours that people need. Sometimes I may slip up and say the wrong thing. That's never happened to you, right? I mean, of course not. You're perfect, right? But, but I'm not perfect. I make mistakes, right? And instead of me having to go back and rerecord that line or rerecord that video, we have an AI version of my voice, and my team can literally just highlight over what I said. If I said, you know, welcome to this podcast episode with Helena, and they wanted to say, welcome to this podcast episode with Jason, they can literally take the word Helena Copy that, replace it, and say Jason, and it will re-say it.

Yeah, with—

yeah, so you can, uh, there's a lot of commercial tools out there. Descript is one of them. They do a great job. Um, there's another— there's several others out there, but we have that for every one of our instructors on payroll. Um, we also, uh, there is new versions coming out in the future over the next 12 to 18 months. You'll probably see this that are gonna be video-based. So you can take a video of Jason and then say, okay, now I need Jason to go teach this thing. Go have him say that. And this works really good if you're using, you know, 30 seconds or less clips. If you start getting into 5, 10, 15 minutes, you can hear the AI is AI, and you could pick out the difference of the two. It's close, but if you know, you know. Like, you could hear the human tone versus the robotic tone. Um, the same thing with video. Um, there's some really cool video stuff that's going on right now. ChatGPT, uh, the OpenAI, the owners of ChatGPT, have something in beta right now. I think it's called Sora.

Sora.

Sora. That's what it is. Yes, Sora.

For video creation.

It looks really, really cool. I mean, it's like I want an establishing shot of a guy snowboarding down a mountain wearing a red hat and blue pants, and there it goes, and it looks like Hollywood production style stuff. So yeah, that's pretty cool and pretty exciting.

I'm excited for that to come out for sure.

And then in my own company, we use AI as part of our exam proctoring service. So if you want to take our exam online at Accolade, instead of being there with a live proctor watching you, we use an AI system, and the AI is verifying you are who you say you are. It checks your ID, it checks your space, it makes sure you're not cheating. And during the exam, if for instance if I was sitting here, some of our exams allow you to have drinks, some don't. If you're taking our short exams, they don't. It's a 30-minute exam, you don't need a drink. And so if you went and picked up a soda can like this, it would actually flag and then alert a moderator for us to come in and analyze the video and say, oh, were you cheating? Were you actually reading the answer off the can, or were you just taking a sip of water, right? Uh, same thing if you have pieces of paper Or anything else. If you're talking in the room, it can know that, it can flag that, it can tell us. And that allows us to do at-scale proctoring with mostly AI technology.

Wow, that's really interesting. And how are you building that? Is it—

So we actually didn't build that ourselves. So we create the certification exams and we create the content. And then our testing provider, which is Certiverse, they have that technology and they're the ones who handle all of the testing. Similar to like if you're doing PMP, Project Management Professional, you take it with Pearson VUE. For us, Accolade creates the certifications. We are PMI in that case. And then Certiverse is like Pearson VUE who does the testing.

Got it, got it, got it. It's, it's very incredible to see, um, the different applications of AI and, and how, how, how, um, it's— and the various different, um, use cases that's being used. I haven't actually seen an AI proctor before, so that's really, really interesting. Um, I know you started your business like 10 years ago. We kind of, you know, we kind of— I almost feel like we grew up Oh yeah. Because we started our businesses roughly around the same time. So turning the conversation to be a little bit about entrepreneurship, for those who are just starting their business, what made you start this business? You know, you are part of the 1% that grew your business to 8 figures and more and got a really good exit. So, you know, in a nutshell, how'd you start your business? How did you eventually, like, find your value proposition to stand out from other training providers? And then, and then, and third part of that question is then how did you manage the process of selling Dion Training and exiting out of that and getting a very big paycheck?

Yeah. So, you know, I wish I could say that I had this all planned out from the start and that I was brilliant and I executed a plan and I knew these were the 20 steps I needed to do to get from $0 to $10+ million. It wasn't anything that cool, really. Really, it wasn't. So when I started my business, it actually was started as a side hustle, right? And actually, if you listen to Side Hustle School and just type in Side Hustle School Jason Dion, you'll see there is two episodes on there. One was when I first started out and got to like ten thousand a month, and the other one was when I got much bigger and was kind of a follow-up episode, like where are they now? And that was I started my company in November of 2016 was when I published my first course. Before that, I was working as a college professor in my spare time, and my day job I worked for the U.S. government, and I was switched. From a job where I was working 9 to 5 for the US government to one where I was traveling a lot. And so what I used to do was I worked 9 to 5, and then at nights I was teaching at the local community college and the local bachelor's degree college near me. During my life, I was switching jobs and I was going to this high-travel job, so I knew I wasn't going to be able to be in the classroom anymore. So I literally turned on a camera and a microphone my last class. I recorded it, which was a Network+ course, and then I took that and put it on YouTube thinking, oh, my students will be able to watch it if they need to follow up. And then people started giving me feedback that they liked the way I taught and they liked it. Somebody told me, hey, you should put this on Udemy. So I put it on Udemy, and then that turned into— I went from $50 a month to $150 to $250 to $500, just kept growing and growing. And within about a year, I got up to about $10,000 per month. I was like, ooh, this is actually turning into kind of a little business. And that's when I turned it into an actual company, which was September of '17, of 2017. And that's when I started hiring a couple of employees. And it was mostly, you know, me and my wife at the time and a couple of outsourcers that we used for the first couple of years. And then around year 4 is when I started actually hiring in other instructors. And around year 6 is when I— it was between year 6 and year 7 is when I sold the company. So yeah, so I mean, it wasn't like a great plan, but it was just one of those things like I started doing and I was teaching the way I teach and people like it and they connected with it. Our success rates were, were number one out there. And so it wasn't like we had a great marketing plan. I didn't have funnels, I didn't have ads. Literally, I just put out good content and hope people would find me. And because in the area I was in, there's only like 2 or 3 other really large instructors in that space. If I could make better content than them and get better pass rates than them, People knew me and then passed my name around. So now if you go and look on Reddit and type in, who should I take for Security+, usually Jason, Mike Myers, or Professor Messer, one of those 3 is gonna come up every single time in one of those 3 orders. Uh, and that's been true for the last 4 or 5 years. And that's really what started making my business go because it was word of mouth and referral, uh, from my successful students telling other people. Uh, the second part of your question was, how do you go about selling the company? Right?

Uh, no, the second part, and that was the third part, is the scaling part. But the second part is like, how'd you turn it from a passion project or side hustle into like a legit, like, big company?

Yeah. So for me, the majority of my sales were on Udemy at first for the first 2 years because that's where I really got my start. And then around my 2-year point, I started doing my own website as well. And so I made a different value proposition on Udemy than my own website. For those who don't know, Udemy is U-D-E-M-Y. They're one of the big online training places, but generally their courses are going to be anywhere from $9.99 up to about $200 for a one-time purchase course. The majority of their courses though are sold between $10 and $50, I would say, because they always have a lot of discounts and things like that. So on the average student, I'd make $3, $5 a student, like not big money, but you do that with 10,000 students a month, it starts turning into something. Right now we do 50,000 to 60,000 students a month, I would say. So we do pretty, pretty good volume there. On my own site, at first Udemy was like 99% of my revenue and my site was 1%. These days Udemy is more about 30%. My site is about 70% because I started scaling up. I started teaching other things and the The offering I have on my site is anywhere from $300 to $1,000 depending on the course you're taking, and we offer more of a high-ticket experience there. So in addition to getting the videos and the practice exams that you get on Udemy, we give them hands-on labs, we give them a textbook, we give them the exam voucher, we give them a pass guarantee and things like that. So that really helped us scale up. But again, we never really did a lot of advertising. It wasn't until after I sold the company that Dion Training started actually paying for advertising and doing advertising out there. In the first 7 years, I spent a total of about $30,000 on ads, and that was because one of my partners who is the one of the backend certifications that we had, they said, hey, because you sold so much stuff, here's $30,000 to spend on ads. Go do it and see if you can turn it into more money. So I never even spent money on ads. Like I was never a big ads person. And honestly, I didn't have really good conversion on those ads because I don't know how to do ads. It was never something I was strong at. So I just created great content and the students kind of shared it and spread it to everybody else. And because I was on Udemy with that very low price offer, it wasn't a high risk for somebody to buy a course for $10 or $15 to see if they like Jason versus going to you know, your own site charging $1,000 or $2,000 for a 1-week course, right? And so I think because my price point was so low and my quality was so high, that really just natively made my stuff go out and spread faster.

And you were the first one to do that. You were, you were an OG, right, on Udemy? For sure, for sure. Okay.

Well, actually, when I got on Udemy, Mike Myers and Total Seminars, that was his company, they were there first and they'd been there for several years and they were number one. Uh, when I got on the platform, it took me about 3 years to become number 1 in all of the categories that we both taught. Um, but the first couple that I did, I was able to take over within the first 3 to 6 months. Um, and one of the keys to that was a lot of times they— the way our certification works that we teach, there's a 6-month overlap between the old version and the new version. So when the new version comes out, you can still take the old version for 6 months. And that company decided they're going to keep the old version out until It ended and then they would switch over. And I decided, well, whenever the new version comes out, I'm going to be first to market. So I had a 6-month head start and that allowed me to chip away from their, their lead and eventually took over. Nowadays they do the same thing we do on day 1 of the exam. Both of us are releasing at the same time and now it's more of a head-to-head fight, but we're still number 1. They're still number 2, which makes me happy. I'm a little competitive, so I like being number 1.

That's really, that's really interesting that you spent no money on ads. I'm still mind-blown that you spent no money on ads and, and did such a Millions in in revenue every single year, so it was a complete word of mouth.

Yeah, I mean it was 99% of our sales comes from word of mouth, right? Now, as as the new company came in, they're trying to take us from where we were at low eight figures to you know eight to maybe even low nine figures. That's where advertisement comes into play. We've now hired a B to B person, a business to business salesperson, and they're building out a team and they're working to expand that because we did some business to business work, but it was like three to five percent, and that was really because. Because, you know, Helena might have taken our course on Udemy and she liked it. So she told her boss and said, hey, everybody should take Jason's course. And then they look at us and go, oh, you guys are much cheaper than everybody else. Yeah, we'll give you the business. And so they reached out to us. Now we're actually reaching out to others, and that's expanding our business even faster as we're, you know, going into hyperscale at this point.

Right, right. Going after the corporate segment instead of just going, going directly to the consumer in B2C settings. Because if you make that one sale, you could sell to—

you're automatically And you might get 50 people instead of one, right? But it is, it's higher contact. There's a longer lead time. There's a lot more work involved with that. It's not as easy just going, hey, IBM, you want to buy 100 courses? Like, good luck finding the decision maker, right?

Right, right, right, right.

Yeah.

A big, yeah, a big challenge with B2B in general is finding those, those decision makers. All right. Now for the third part of the question, which is talking about the exit. Yeah, that part really intrigues me.

Yeah. So for me, I had gotten offers over the years, even as early as year 3 of my business, I had people reaching out going, hey, we're interested in maybe buying you. And we would do some talks with them. And, you know, I was never really looking to sell necessarily because this was something I, I never built with the idea of selling this company. Whereas my new company, we have built this from the beginning with the idea that we will end up selling this one day or we will continue to own it and run it. But either way, we're set up from the beginning that we can sell it. I would tell your audience, if you want to do an exit one day, try not to make the brand about you. When I first started teaching, my website was JasonDeon.com. And right now, if you go to JasonDeon.com, it just redirects to Deon Training. After about 2 years ago, oh, you know what? What if I want to have another instructor like Helena teaching project management or AI? Well, it's kind of weird to have Helena teaching on JasonDeon.com, right?

Yeah.

So we decided to become Deon Training because I'm so creative with names. Again, it's very much a personal brand. And at the time we made that decision because my name of Jason Deon was pretty popular in my space. And so we wanted to make sure Deon was still in the title because people search for either Jason Deon or Deon Training or Jason Deon Training to try to find us. But in general, I would recommend finding a short, catchy name. I would find something that could be trademarkable. I would find something that you can pass off that is not necessarily tied to you. And if I was making a software as a service AI company, I would not call it Deon Software, right? I would call it something else, right? For instance, when I started the certification brand with my partner, it's not called Deon Certification, right? And we didn't do these certifications as part of Deon Training. We made it a separate organization. It's a separate company, and we have a name that is short. It's 7 letters, Accolade. Easy to say, harder to spell. And the reason it is harder to spell is because we wanted the .com. We wanted something that was trademarkable. And if you have the word accolade spelled the right way, A-C-C-O-L-A-D-E, that's not trademarkable because it's just a dictionary word. But because we have accolade spelled a little different, A-K-Y-L-A-D-E, we can trademark that. We also are able to get all the social media handles for that and all those other things that go into it. And that is now brand equity that when we go to sell, we get a valuation on that too. When it comes to selling, definitely, definitely, definitely get yourself a good business broker who knows what they're doing and has done these deals before. When I went to sell my company, the business broker I used, James, he's fantastic. He knows a lot about what he's doing. But like I said, he only takes on 5 to 10 people a year because every deal is really, really time consuming. It's a, you know, 3, sometimes 6 month process. And he's told me that up to 90% of deals in the acquisition side fall through at the last minute. And so his company, they only pick deals that they're pretty confident are going to get to the finish line. And he was able to get us to the finish line on our deal. The other thing is, if you go without one, it's of like selling your home without a real estate agent. You can do it, but you'll probably lose a lot of money. The first offer I had versus the offer I got when I went to sell, um, it was almost half of what I ended up ultimately selling it for.

Wow.

It was about 40% discounted off what I was able to get during negotiations using my business broker. And that negotiation happened over about 48 hours. Um, literally when I talked to my business broker, said, I've got this company interested in me, here's what they're offering. He called them up, And within 24 hours, we had a raise of probably 25%, 30% over what they originally offered me because now he was involved. And we gave him— I gave him another 24 hours and he was able to get it up, you know, even more, another, another 15%, 20%. So it was about 40% higher than I would have been able to get on my own. So it was, it was a significant difference, right?

Wow.

And so, you know, same thing when you, you know, buy a house without a real estate agent. Often you'll be like, oh, well, Helena is selling her house without a real estate agent. I know you're not paying the 6% real estate commission. So guess what? I'm going to ask you for a 5% or 6% discount because as the buyer, I want less money to pay, right? And so they're just going to try to get— their job is to get for the lowest price possible. Your job is to sell it for the most money possible. And somewhere in the middle, you guys come to a number that you both can agree to. The other thing when it comes to that is expect to pay quite a bit of money when it comes to selling. It is not free to sell your company. You need lawyers, you need accountants, you need financial forecasts, you need a business broker. And all those people want to get paid. I can tell you that we spent high 6, low 7 figures between lawyers, accountants, business brokers, etc., to get this sale done.

Right.

And that came out of my pocket. And whether we closed the deal or not, I still had to pay those folks. Right. The business broker, he worked on a commission still that only got paid once the deal closed. So he's incentivized to make sure the deal closes. But the lawyers and the accountants, if the deal fell through, I still would have been paying them $80,000 or $100,000 in fees. So it's not a cheap thing to do. And that's why most people, when they go to look at buying you, they're going to want to see revenues somewhere between $5 and $10 million and net profits of $1 to $5 million. If you're under those limits, you're probably too small to get a decent offer from a good enough company. When I first started getting offers, it was like, we'll give you $1 million or we'll give you $2 million. And it was so small that like finding a lawyer or an accountant who wants to work with you isn't really going to work either, right? Because the other side, the person who's buying you, they have lawyers and accountants they have to pay too. And so they're investing heavily into this deal. And why would they spend spend $50,000 or $100,000 or $200,000 or $300,000 on lawyers when they're going to do a million-dollar purchase. It's just not a big enough spread to make it worthwhile.

Yeah. So 2 questions there. One, what kind of due diligence are they doing? Sorry, forgive my naiveness. I've never sold a company before. And then the second part of that question is, how do you know that the company that you want to sell to is legit? Because I've had a few offers, but it was really just competitors trying to see what my revenue is and how I'm running my business. So How do you— so how do you know they're, they're actually, you know, the right kind of people who want to do the deal and not just competitors trying to see like what you're doing inside of your business?

Yeah. So I'll start with the second question first, which is how do you know they're legit? And then we'll talk about what due diligence looks like. And the first part is you're going to do a little bit of due diligence on them when they first reach out to you, right? So when I ultimately went to sell my business, as I said, I had multiple people over the years and some of them are just tire kickers, right? I've had some that come in, they kick the tires and they're like, yeah, what are your revenues? What are you doing? Blah, blah, blah. And you tell them and they're like, okay, we'll give you 2x your revenue. And I'm like, go away, right? Like, I could stop filming courses today and I'll still make 3 or 4x on my revenue over the next 3 years just because naturally I have a big enough audience. So if you're talking 2 or 3x, it's just not even interesting to me whether you're legitimate or not, right? So that's one thing you look at as far as are they legitimate. One is you want to talk to the— when you're talking to them, say, who have you bought previously? Tell me some of the names of the company. Do you have points of contact I can talk with? So the company that ultimately bought us, as we got into those discussions and we had some NDAs in place, they said, hey, Hey, if you want to learn about us and what it looks like as a founder when you sell to us, here are our 6 other founders. Call them, see if they want to talk to you. They'll tell you the good, the bad, and the ugly, right? And I can tell you that I've talked with probably 5 or 6 founders since when I sold my company. Last May is when I sold my company. And new people, like if the company who bought us wanted to buy Helena's company, she may call me up and say, hey, tell me what your experience was. And I'll tell her, this was good, this was bad. Yes, they're legitimate. Here's the things you should work out for. You know, You know, when during the transition, when they buy you, here's what it's going to look like. Here's what it looks like as a founder still working in the company after you don't own it anymore, because that's a whole different shift, right? You don't own the company. You don't make the decision. Somebody else is because you now have a corporate overlord, right? Because it's their money, not your money, right? And so those are all different things too. So that's that's a big piece of that. The other thing is you can look at the companies, and there are websites out there where you can look at private equity firms and see how much their holdings are, and are they real? Are they legitimate? The other thing is, what kind of business model do they have? Everybody who buys businesses does it under a different model. For instance, the people who bought us, their model is they invest a lot of money into the company because they want to take it from where it is and get to 5 or 10x what they paid for it, and then they're going to sell it off to somebody else, right? That's their goal. And so they want to do value add. And so for me, that was a good fit because I want the business to keep growing. Some other companies that approached us, when I started looking into them and I started looking at what they did, and as As I looked, a lot of them, they would go in and go, oh, you've got 30 employees. We think you can do the work with 20. Oh, you're spending all this money on, I don't know, Zoom. You guys should just use Google Meets because it's free. And they just want to do cost cutting, right? And that's fine, but it can also hurt your business, right? That can lower your quality, right? But that's an instant way for private equity firms to go, we paid $10 million for this business, and if we can cut off 20% of their costs by firing Americans and hiring people overseas or whatever it is, we can increase our profit ratio really quickly, right? In our case, our profit ratio was already good. You know, our profit margin was really high already. And the company that bought us ultimately was, was a much better fit for that. I had 3 companies at the end that we were talking to. One was one that we just realized was not going to be a good cultural fit, even though they offered me more money than the other 2. One came in low and they were more of the, we like to come in and slash budgets. And one was the one I ultimately sold to, which was a growth company, a social good company. They're a really big firm with a lot of money behind them. And they brought in really good people to help us go to the next level. The one that wasn't really a good fit is just knowing yourself and knowing your company. At our company, we are big on DEO, so we have a lot of people in our company that are Black, white, Asian, gay, straight, trans. It doesn't matter to us. We like all of that, right? We want good people and we don't really care what you do in your personal life. Just do good things for us and our customers and take care of people, right? One of the companies that wanted to buy us, their background was buying and selling content and processing and technology around churches. And so they were a religious right organization. And I knew that that was not gonna be a good fit when, you know, 3 or 4 people on my team are gay, lesbian, trans, bisexual, whatever. And we had just a— it was gonna be a culture shock, right, between us and them. And so it's like, this just isn't going to be a good fit, even if they wanted to pay more. It just wasn't who I wanted to be with, right. And so we said no. And so ultimately, we chose who we chose.

Mm-hmm. Yeah, that makes sense. And then going back to the first part of that question, which was about what that financial due diligence Due diligence looks like and why, um, you know, you had to pay $100 grand to lawyers and accountants, um, for that sale.

Yeah, so the due diligence, right? Yeah, so, uh, on the due diligence side, what it's going to look like is we're doing due diligence on them and they're doing due diligence on us. For our side, it's a lot less, right? We're looking and saying, do they have the money to pay me, right? How are they going to structure this? Where is the money coming from? Because we had one offer where they're like, hey, we'll pay you all this extra money, but we're only going to give you 20% of it up front. We want you to give us a loan over 5 years for the next 50%, and the other 30% we're going to give you as a commission based on profits over the next 5 years. And it's just a really structured, weird way of doing it, um, and we just weren't comfortable with that, right? Because now I'm doing an owner's finance loan, they're using my money to pay me back, and if they drive the business to zero, I lose both the business and the money that I lent them. And I just wasn't comfortable with that, right? Um, the company that bought me, they had— they're owned by, you know, we are— they have a big private equity firm that has multiple smaller firms underneath them. And the one we're under is the e-learning pod. Uh, and that one, the training and development pod, has, you know, 5 to 10 companies in it. The company that owns that has $6 billion behind them. They're a huge, huge, um, you know, investment, uh, house. And, and they've done really well over the years. They have a really strong backing. They've been around for 15, 20 years. I've talked to a lot of their founders from their portfolios, and I was comfortable with that. So that's kind of my due diligence on their side. The other side is their due diligence on us. So they brought in accountants, right? And when I started my business, this was me. This wasn't me thinking I was going to sell one day. And if you're going to sell one day, I would encourage you to use accrual-based accounting. If you don't know what I'm talking about, ask your accountant and look at getting your books switched over to accrual-based accounting, because anybody who wants to buy you is going to use accrual-based accounting. Most of us as small business owners, we don't use accrual-based accounting. We use cash-based accounting. And the difference is, for instance, Udemy. Udemy, let's say I have $100,000 that I earned in February. Well, I don't actually get the check for that until April 7th. It usually runs about 35 to 40 days behind. The month has to close out. They give me a 30-day return window and then they send out the check, right? And so for us, from the time I earn the money to the time I actually get to account for it is about 45 days in cash-based accounting. Because in cash-based accounting, you get credit for it when it hits your books, when the money hits your account. With accrual accounting, based, you actually take credit for it today. So if I do a big B2B sale and I'm going to sell $300,000 worth of courses to Helena's company, when she signs the purchase order, that counts as money today in accrual-based. It doesn't count as money until I actually get the payment, which is 2 months from now, right? And so it's just a difference in the way the books are done. And because we were not in accrual-based method, one of the things they had to do is they literally had this accounting firm come in and redo all of our books for the last 3 to 5 years And say, here's what Jason says our P&L is. Here's what our profit loss is, and here's our balance sheet. Now, what does it look like if we did accrual? And when they did that, they actually found, oh, his business is actually worth more than we thought. Right? And so that was that was one thing we could do. So that's one of the pieces. The other piece is the lawyers. Right? The lawyers are going to come in. They're going to look at: Do you actually own the content you say you do? Do you own the IP you say you do? So you have an AI company, Helena. You're building algorithms. You're building you know large language models and all that stuff. If that's all technology that you own. That is something that's going to give your brand more equity, and so they want to verify you do own it, and it wasn't your you know old business partner who did it or some subcontractor who didn't give you the right rights. Do you own the trademark on your name? Do you own the color schemes and the branding and the logos? All that stuff is stuff they're checking. If you have agreements with other people, for instance, we had several contracts with folks like the Department of Defense, the U.S. Air Force, the United Nations. We have some other large companies like SAIC and ServiceNow and other people like that, and they had to verify. Are those contracts valid? Are they going to continue on after Jason sells the company, or are they going to leave when Jason leaves? Right. And that's all things that they're looking at to verify the numbers I told them. Because as you're doing your initial conversation before you enter into an agreement, you're saying, yeah, my revenue is about $5.8 million a year, my net profits are $2.3 blah, blah, blah. And they make you an offer based on that. But then the due diligence is, let's verify what you said is actually true. And in our case, it was fairly But at the end of the day, once they did all the conversion over, we were off by like $200,000 on this multimillion-dollar deal. And so at the end of the day, instead of rewriting the whole deal and going back and forth, we just chose to give Jason a lower salary for 2 years to account for that. So instead of making $200,000 a year like I was paying myself before, now I'm making $50,000 or $60,000 a year to overcome that difference over the 18 months until it's equal. And then I go back to a regular salary. So there's lots of ways you can structure it. And these things can be very complicated, which is why you need a lawyer. They could buy your business. They can buy your assets or they could buy a part of your business or all of your business. In our case, they didn't want to buy the whole business. They wanted to buy our assets. And essentially what we did was we sold all the assets, which is all the contents, my name and likeness, the domain names, the IP, everything we have. And then the structure that we had, the old company, we just simply shut it down. And the reason they do that is twofold. One is for them, they're only getting the good stuff without any kind of debt or liability. Because if somebody comes back and says, hey, Jason sold us a thing, you know, 6 months ago, Dion Training Florida sold us this thing 6 months ago, you now owe us a couple hundred thousand dollars. They go, not our problem, go talk to Jason, right? That was his company. Um, but if they want new content, they, they get that from the new company that now has that, and that's the new liabilities going forward. And that's a very, uh, common way of doing it. It's called an asset purchase agreement. The other good thing about doing that from a tax side as an American, um, they were buying things like my brand name and my goodwill, and that's been earned in the industry, and a lot of our sale was tied That, that is taxed as capital gains, not as ordinary income, where if they just wrote me a check for the whole company, it would've been ordinary income, which is taxed at 37%. Capital gains is 15 to 20%. So you're paying half as much taxes based on how you structure your deal. And again, that's why paying $20,000 or $30,000 to an accountant who's going to save you 20% on a multimillion-dollar deal, well worth it, right?

Right.

And that's why I say you need these lawyers, you need these accountants because they'll make sure you don't make stupid mistakes. Don't be You know, pennywise and pound foolish, right? Is what it comes down to. Sometimes it pays to have these type of folks in your corner. And the other thing is the company that's going to buy you, they've got a team of accountants, they've got a team of lawyers. And if you don't, they're just going to run over you and you're going to sign yourself up for stuff you don't know what you're even signing yourself up for. So every document was going back and forth and we had things like, can Jason teach elsewhere? Can Jason go on podcasts like this one? Right. In the agreement, there was things that says Jason can't go host his own podcast. Right. But we have that I'm allowed to be a guest because I'm talking about my experience or talking about my companies. That's okay. I'm not allowed to go start another training company called Jason Training to compete with them for 5 years because we have a non-compete in there. Right. And there are certain things. And because I had this, I had a couple other companies, we had to go back and forth and say, well, yeah, I get that and I'm not going to do training, but I do have this certification company and I'm not going to shut that down because you guys want to buy my other company. So we had to have exceptions. And it says, okay, Jason cannot do training for that company, but he can sell certifications. And if he wants to do training, he's going to do that under Dion Training because they're one of our partners that teaches those certifications. So there's a little bit of this going on. too. But all those things that can be negotiated, right?

Yeah, yeah. Very, very interesting. I love talking to you about this stuff. Um, and I know we had a lot of conversations about this offline as well. Um, oh yeah, I know you are definitely an early adopter of any kind of technology, um, AI in particular. Um, so, um, how do you motivate your team and really get a culture in your company around, um, embracing technology and embracing the latest trend and And then how do you get your team to continuously learn? Because, as you know, change is a very hard thing.

Yeah, change is really hard. I think it goes down to the culture you build in your organization upfront and continue on. If you are a company that has been around for 100 years, like let's say you are, I don't know, IBM, they've been around for decades and maybe even 100 years at this point, right? It's hard to steer that ship. We've been around only six or seven years by the time we sold, right? And so it's much easier for me to create a culture from the beginning that supports that. I have always been. We are a technology company, right? We teach technology and we use technology, and we even build our own technology because we have our own developers. And so for us, we already have that that mentality in our workforce, so it's not hard. And I think it's a matter of hiring to the the what you want in your organization. So for us, we want to be able to be agile. We want to be able to move fast. We want to be able to try new things out. And I think the other thing is we've created a culture where it's. Not a zero defect mentality. I don't know if you ever worked for a company where you have a zero defect mentality. You make one mistake, it's like you're fired, right? And when you do that, people just stop taking risks. And new technology is a risk, right? I am a very risk accepting person. I have a high risk tolerance for things like that. Otherwise, I wouldn't have been an entrepreneur, right? And so for us, it was something that has gone through our company from the beginning. We've always had the philosophy of we'll try things. It's non-critical deaths. And so if something does Does it work? We just roll back and do something different, right? We play a lot with ChatGPT. We play a lot with AI technology. We've created a system where we can write our own practice exams using AI technology, and I'll tell you even that it's still 70% good. Like there's still 30% of stuff we have to change, and so we've implemented processes where we have humans in the loop. So once the thing comes out of the AI system, a human reads it and says yes, it's good, or no, it's not. If it's not, we go and redo it. And so I think having those checkpoints allows us. us to take calculated risks without affecting our students negatively. So, um, Mm-hmm. I think it really just comes down to a culture of making sure you have psychological safety, which means you don't tear people down for making mistakes. We learn from our mistakes and we move forward. We try new technologies, and if they don't work, it's okay. We just go back to the old technologies, right? Um, anytime something new comes out, we will try it out usually in a test environment first with some subset of our employees, and then we'll go from there. Uh, for example, when ChatGPT came out, you know, 3 or 4 of us out of my 30-person team got licenses. We play with it, we use it, we learn the best way to use it, and then we share that with the rest of the team, and then we give them all licenses. So we have our early adopters who are a little more tech advanced, and then we bring the rest of the team along with us and say— and the other big thing is telling them how it's going to benefit them, right? It's not about what it's going to do for the company and our bottom line. But if I said, Helena, you're currently spending 8 hours a day reading student reviews and writing up a summary for Jason to know, are the students liking his courses or not, hey, if you do this with AI, we can knock that down to an hour. And we've done things like that, right? And so it's like, look, I'm saving you time. Just tell me what you need. And one of the things we do is once a quarter, everyone in our team goes and logs their time of everything they do for 2 weeks. And that goes all to our COO. And she looks through it and says, hmm, this person here is spending 8 hours a day doing reviews. Let's see if we can do something to automate that or machine learning that or AI that to get it out of our way or at least reduce it. And so that's one of the ways that we continually look for new opportunities. opportunities where we can save time, save money, and make our employees' lives better. Because if you're copying and pasting data 100 times, that's kind of a waste of time, right? I want to use your brain. I'm paying you to use your brain, not to copy and paste stuff. I can write automation to do that.

Right, right, right, right. And, and, you know, I even have an example of that. Like, you try GoHighLevel, and we realized it wasn't the best fit for your company, and then, and then you just moved on to another tool.

Right? And there's nothing wrong with GoHighLevel. I'll put that out there because I know you, you like GoHighLevel as well. It just didn't have all the features that we needed. It had a lot of— it had a ton of features, but it was— I felt like it was much more focused to people who are doing marketing and selling, and that wasn't what we needed in our business. And so we've chosen a different tech stack that we've gone with that has multiple different pieces that we've interconnected. But also that's because we have a bunch of developers, so we can build our own stuff. And, you know, if you don't have that, GoHighLevel may be a great option for other people.

Right. I was just giving that as an example of where you try something with a small set of your team. So we did that with you, Pip, and your sister, Allison, and then didn't, and it didn't work out. So then you move on to something else. And, and so this kind of brings me to one of the last questions, last, second last question, I'd say. How do you keep your emotions in check when you're working in an industry, starting a business that's very volatile? And I think mindset is so important, but it's so overlooked as well.

Yeah, I think mindset's really important. I think one of the things that really worked well for me was I started this company as a side hustle. So it wasn't that I needed this company to work to make money. I already had a really good job making, you know, $100,000, $200,000 a year working in cybersecurity. So I was making good money. And so that was paying my bills. And then I was working on this company nights and weekends. And once the company got big enough, I no longer needed to do that. And I am a little risk averse on the financial side. So like by the time I quit my day job, I had already like 3 or 4x'd my day job's income. I've seen other people that kind of get close enough to like, hey, I'm making $100,000 a year, I'm making $50,000 in my side business. If I quit my day job, I can get this to 100 or 150, and they'll do that. But if it doesn't work out, now you're kind of screwed because you need 100 to live, but you're only making 50. In my case, uh, we— I didn't quit my job until I had, you know, 2, 3, 4x my annual income on the side business, and it had been going for 2 or 3 years before I actually quit my job and went there full-time. So for me, it was, it was a risk tolerance thing there. Um, the other thing is, in my personal life, I had already gotten my house paid off, my cars paid off. I owed nobody anything, and I had a large emergency fund. So if the company failed. I also have a very marketable skill set. I know if I wanted to get a new job, I could put on LinkedIn today, Jason's looking for a new job in cybersecurity, and I guarantee I would have at least 100 offers by Friday. It's just like my name's known, people know who I am, and my skill set is very valuable and highly desirable. So it was, it was less scary for me under those circumstances.

Mm-hmm.

So I think the big thing is get yourself out of debt, get yourself into a good financial shape, because that will take a lot of the pressure off. The other thing is set up good boundaries in your life. I did not do a good job of this, right? Because I was doing a full-time job and then I was working on this nights and weekends. So often my wife and my kids got the short end of the stick because I was working and like, oh, I can't do that tonight, I'm working. Or I'd be working till, you know, 2 in the morning and then waking up at 5 or 6 and doing it again because I had to go off to work from 9 to 5. And then when I got home, I'd play with the kids for 2 or 3 hours. When they went to bed at 7 or 8, I would then go in my office and work for, you know, 5 or 6 more hours. So I think it's important to take care of your mental health and your stability. There, otherwise it will show. People can sense desperation in your sales. They can sense it in your voice when you're doing meetings with clients. So get yourself into a position where you are coming from a position of strength as you start your business, and do it as a side hustle if you need to first. Get yourself built up, and then you can go full time into it. I think that really helps.

Yeah, for sure, for sure. I look back on my own journey, and I think whenever I was desperate, I didn't get the sale. But whenever I'm just like, c'est la vie. Whatever happens, happens. Yeah.

I have a course on how people can get jobs in cybersecurity. And one of the things I talk about in there is it's the same kind of thing, right? Employers and hiring managers, we can sense desperation, right? When you don't have a job and you're looking for a job, you're almost like, please, please, please hire me. And you're not saying that, but we can read it on your body language. Whereas if you already have a job and, and, and they're like, hey, would you like to work for us? Like, eh, I can take it or leave it, right? It just, it, it hits different. And I equate this to like back in high school, you know, I, you know, there's all these girls I, I was in high school. I'm like, ooh, I'd really like to go out with that girl or that girl, and I would ask them out and they're all like, no, no, no, no. I'm just getting turned down left and right. Nobody wanted to go out with me. Then I got a girl that I was going out with, and as soon as I had a girlfriend, like a month later, all of a sudden all these girls start flirting with me and they all want to go out with me. I'm like, a month ago you guys couldn't care less, but now that I have a girlfriend, now you want me? Apparently it's like, oh, well, if somebody wants you, then you must be valuable and there's something there that we're not seeing. But when you're single, nobody wants you. I think it's the same kind of thing in business, and it's the same kind of thing when you're trying to make It's like, especially when you're having to do the outreach and you're cold calling people, it just to me, I can tell you I hate when people cold call me. I hate when people send me emails offering their services. I'm like, I didn't ask for you. I don't need it. Don't fill up my inbox. And you're already coming from a position where I just feel like you're desperate. If you're having to come out to me and tell me about your web design services or your application design services, I'm like, what told you that I was even looking for a website? Go look at my website. It's beautiful. I have a whole development team. I don't need you to. So you're obviously just missing the mark because you're sending out to millions of people. And to me, it's just you're wasting my time and you're desperate. And that's just not a good way. Sure, sure.

Of course. So this brings me to my last question, but it's a 2-part question. For those who want to explore entrepreneurship opportunities, especially with the rise of how easy it is to now start software with like no-code tools like Bubble, what kind of entrepreneurship opportunities are you seeing that emerge between like for AI-driven cybersecurity? related, uh, fields that hasn't been, been explored or, or is, um, underexplored that could, that could turn?

Um, so, you know, this is one of the big challenges, right? If you're starting out as a small person, like a 1 to 5 person team, um, AI is changing so fast and there is so much money being thrown at it through Silicon Valley that you will not be able to move fast enough to make a dent. Um, and so you're going to do a lot better doing things in a very small niche niche environment. What I mean by this is I actually had a company that I started and actually closed it down earlier today. I signed the papers to close it down. I started a year ago as a side project, but honestly, I just didn't have enough time to dedicate to it. And I had a very small team. It was me and 2 other partners, and we were the developers. And we had this great idea for a product that we wanted to build using AI. The problem is, because we could only dedicate 10 or 20 or 30 or 40 hours a week to it, and we only had 3 people, it was going to take us a year to build this thing. And AI is changing so fast that as we were building it, ChatGPT 4 came out and it already had like most of the capability that we would need, right? And 5 is going to come out soon, right? And so if you don't have a big enough team and enough money, this world is changing so fast that if it's going to be— if you have a year-long pipeline to get to development, to get to, you know, production release where you can start making money, you might as well not even— yeah, you're obsolete by the time you get out, right?

Yeah.

And so that's what happened with us. And so I was like, you know, honestly, I'm just not going to throw good money after bad in this case. And we shut down the company and said it just wasn't worth it. We had thought about doing a pivot to a very niche area And that's the way that I think you can sustain this is you're never going to be a billion-dollar company that way. But if you're looking for a good, you know, $1 million, $5 million, $10 million a year business, you can do that by serving a small niche in a very focused way that other people are overlooking. For example, you know, one of the things that we— that I have a background in is working with lawyers. And lawyers have a lot of work that they do that everything is kind of form-based for a lot of the stuff they do, but each form has to be kind of individualized to that person based on their needs. Well, instead of hiring a paralegal for $50,000 or $100,000 a year to do that, you could create an AI system that you could maybe sell it to them for $50,000 a year as a license, and that can replace 3 of their paralegals because now the AI system is doing a lot of the drafting for them. That is a no-brainer for a lawyer to say, hey, I can get rid of $150,000 in labor if I pay you $50,000 for software. Sure, I'll do that.

Right.

And it makes their job better. faster, cleaner, it's repeatable, yada, yada, yada, right? So if you can identify things like that, that's where you'll be able to make the biggest impact. I think it's going to be challenging to do something in cybersecurity at scale because there are so many people already investing heavily into it with a lot of money. You know, tens of millions of dollars is going into this area. And I think really the big areas that we see in cybersecurity is defense and automation. So using zero trust networks that AI is going to look at your system and determine, is it safe enough to add to the network? And if it isn't, how do we adjust those settings to make it safe before we allow it to join the network. We also have AI attack. So if we're doing pen testing, we can create AI tools that think like an attacker, and we can configure them to say, act like a script kiddie, or act like a hacktivist, or act like a nation state. And those are, you know, kind of low, medium, high level of skill. And then attack people's networks using that on a contract basis. And if we could train AI to do that, that's a lot cheaper than having somebody who's making $100,000, $200,000, $300,000 a year doing it. Now, does that mean you don't need that pen tester? No, it doesn't. It means that that person now doesn't have to do things like port scans and the easy stuff because you can AI that out and then use them for the high-level quantitative, uh, and qualitative analysis stuff. So that's kind of where I look at this stuff. Um, and then the other thing I would say that's another really great business opportunity is where you're making custom AI tools for small and medium-sized businesses. A lot of times small and medium-sized businesses are overlooked, and that's an area where you can really provide a lot of value because they don't have millions of dollars to hire a full development team, but they may have, you know, $100,000 a year to hire you and your coders where you're now sharing that across 5, 10, 15, 20 different, you know, companies that you're supporting at a licensing fee of $10,000 per month. I have a friend who does a lot of ChatGPT individual tools using the new extensions and plugins. He custom develops those for organizations and he charges a $10,000 per month retainer. So it's $120,000 a year. But for that, he will build all the customized automations and AI tools that you need for your business up to a certain amount per month and a certain amount amount of coding hours and all that. But it's a lot cheaper than hiring a team of 10, and he's got a team of 10 or 20 people, and he spreads them across different teams and different companies to support them. So it's more of that shared services model in that case.

Sure. Yeah, absolutely. And that's— and, and what you said was like one thing that I really thought about when I was thinking about how do I play in the AI space. I knew I wasn't going to be able to compete with like the Microsofts and the, and, and the, you know, tech giants of the world. Um, so I was like, I'm not even going to bother, like, create my own language model. We're trying to compete with that world because Facebook can just outspend me any day, any day of the week. So then when I thought about how do I compete in this landscape, I turned to services because even though those tools are available, my thinking was a lot of business owners are just not going to want to do this work.

Yeah. And I think there's, you know, 3 different models there, right? One is training them how to use these things so they could do it themselves. One is done with you where you come along alongside and you're almost coaching them through like, hey, let me give you an hour to teach you how to build this chatbot, blah, blah, blah. We set it up for you. Now you, business owner, need to go and configure your specific prompts. And then the third way is what we call done for you, where you basically charge a lot more money, but you're going to sit down and interview me as a business owner and say, Jason, what do you need? Oh, I need, I need an AI tool that does this, this, this, this. You go, great. You go off and build it. You do all the stuff and then you just turn it over to me and I pay you a licensing fee. Right. And those are kind of the 3 different ways that I see for us small and medium-sized folks that we can get involved in this. And there's a lot of opportunities there. You'll still make a lot of money and make a really big impact, but you're not going to have a billion-dollar exit with a services company, right? When you're trying to sell a services company, you're looking at 1 to 3x on a sale. If you have a technology company, you're probably looking at 5 to 8x on a sale. And so if you're doing $1 million a year in net profit, that is the difference between $1 and $3 million versus $5 to $8 million. So, you know, more than double or triple as much if you have a technology that you can brand, whereas services, it's an individual one-off contract each year. time.

Sure, absolutely, absolutely. Well, Jason, thank you so much for being on the podcast and sharing all your wisdom with us. It's always a pleasure to speak with you.

Definitely. Thanks for having me, and happy to come back anytime you need.